How my OpenAI scholarship inspired me to create my own product

The origins of Why I care about error reporting and user journey

I dreamed of being a professional ballerina since I was 3 years old. I loved the idea of being able to dance on my toes, light as a feather. My dream came true for a while. I graduated from a professional ballet school where I studied for 10 years and danced with international ballet companies.

In my early twenties, I started experiencing a sharp pain in my knee. At first I didn’t pay attention to it. Over time the pain became more intense and I could no longer ignore it. Long story short: I quit dancing because of this injury.

Teaching and choreography didn’t appeal to me. So I decided to do something else in life. At that time I didn’t know what I wanted to do because my whole life was centered around dancing. I had no experience of anything else. I started experimenting and trying new things. I graduated from two universities, with degrees in economics and journalism, and tried many professions. I was just trying to find myself. The more I searched, the more lost I felt.

It was a dark time and I was burdened with existential questions: Who am I? Why do I exist? What should I do with my life? When I die, what contribution will I have made to humanity? It took me hours of therapy, healing medicine, and years of meditation to realize that all of those vague existential questions ultimately don’t lead anywhere. They arise from our ego, and if we cut out all of those superficial questions that we create in our mind, we come to realize that there is literally nothing. Reality is happening in our mind, nowhere else.

My scholarship at OpenAI

While I was in this, “who am I and why am I here” mode, the AI hype of 2015 happened, when everyone was talking about AI, neural networks, and kept dropping confusing buzzwords that I didn’t understand.

The term ‘hype’ has a bad connotation, but it does have some positive aspects. I genuinely fell in love with AI. I realized that I wanted to make a significant contribution to AI technology. Just the idea that people can potentially create an absolutely new form of intelligence fascinated me. Think of your cat: it’s not aware of bitcoin, COVID, or anything else people worry about. Imagine how many things are happening in the universe that we don’t know about.

With AI as a new form of intelligence, we can discover new laws of physics, colonize Mars with robots to build infrastructure for us, and make progress in other directions. It’s truly fascinating. I wanted to be a part of that movement.

The only way I could see for me to meaningfully contribute was to gain expertise in AI and to band together with like-minded individuals. That motivated me to start coding and learning data science.

I am self-taught. You can see some of my data science projects here. Eventually, I ended up going through OpenAI’s scholarship program.

During the program, I further developed my skills in machine learning. While I used different libraries for my research projects, I ran into various issues when I used ML libraries. GitHub issues and Stack Overflow were not enough to resolve my technical issues. 

I wanted there to be a product that could identify the crashes I faced while using libraries. Something that could send a crash report directly to the creators of the software so that they could analyze the problem, fix the issue, and then inform me of the solution.

This idea stuck with me for a while. Eventually, I decided to take the bull by the horns and create this product myself.

Creation of Bugout

This is how originated. When I shared my frustrations with Neeraj, my co-founder, he shared another side of the problem. He used to work at Google on the TensorFlow team. It was really hard for them to understand the issues TensorFlow users had, especially Windows users, who were severely under served. He saw the problem from another angle as a library maintainer. He agreed with me that the problem exists and that there needs to be a solution.

So, together, we created to help maintainers of APIs, open source libraries, command line tools and any software creators who license their software to customers collect error reports and usage metrics to better understand how to improve their software and deliver much better customer service.

But, more importantly, we built the product to help end users like me when I was at OpenAI. We catch errors and make the fixes available in a public knowledge base that never goes stale.

Bugout is our way to support builders!